A new objective routine for the identification of individual blocking events has been developed which is applicable to data sets of different horizontal resolution. It is based on the 500 hPa geopotential height anomalies with respect to a mean seasonal cycle. Three features of blocking are checked with this method: the intensity of the anomaly exceeding some threshold, a certain extension prescribed by a peripheral value of geopotential height, and a duration of at least 5 days. The routine is applied to ECMWF analyses and to climate simulations of the Northern Hemisphere, which were performed with the atmosphere model ECHAM3. In winter, the blocking statistics derived from the observations are in accord with the results of other identification techniques. In other seasons the differences are larger due to a different definition of blocking. It is shown that the model ECHAM3 simulates the observed two‐dimensional frequency distribution of blocking events quite realistically.